🔬 mesoscale physics

Reinterpreting Memory Effects in Nonequilibrium Systems: From Temporal Dynamics to Steady-State Signatures via NEGF

This paper investigates memory effects in two-dimensional nonequilibrium lattice systems using the NEGF and Schwinger-Keldysh frameworks to demonstrate how distinct scattering mechanisms (static disorder versus electron-phonon coupling) generate Markovian and Non-Markovian dynamics, respectively, which are identifiable through spectral function signatures and analyzed across various microscopic models.

Original authors: Pragya Chaudhary

Published 2026-05-29

📖 5 min read🧠 Deep dive

CC BY 4.0

Original authors: Pragya Chaudhary

Original paper licensed under CC BY 4.0 (http://creativecommons.org/licenses/by/4.0/). ✨ This is an AI-generated explanation of the paper below. It is not written or endorsed by the authors. For technical accuracy, refer to the original paper. Read full disclaimer

The Big Picture: How Systems "Remember" Their Past

Imagine you are walking through a crowded hallway.

Scenario A (Markovian): You bump into someone, they push you, and you immediately forget the bump. You keep walking as if nothing happened. Your next step depends only on where you are right now, not on the bump you had five seconds ago. This is called Markovian behavior (no memory).
Scenario B (Non-Markovian): You bump into someone, but instead of just pushing you, they grab your arm and swing you around. You feel the effect of that push for a long time, wobbling and adjusting your path based on that interaction. Your next step depends heavily on what happened to you in the past. This is Non-Markovian behavior (with memory).

This paper is a theoretical study by Pragya Chaudhary that investigates how electrons moving through tiny, 2D materials (like a flat grid of atoms) behave in these two scenarios. The author wants to know: Does the electron forget its past instantly, or does it carry a "memory" of its interactions?

The Two Main Characters: Static Noise vs. Dancing Phonons

The paper looks at two different ways electrons get "bumped" or scattered:

Static Disorder (The "Static Noise"): Imagine the hallway floor has random, stationary bumps (like pebbles). When an electron hits a pebble, it bounces off. It doesn't lose energy; it just changes direction.
- The Paper's Finding: This is like Scenario A. The electron forgets the collision almost instantly. The "memory" of the crash disappears so fast that the electron behaves as if it has no memory at all. The paper calls this Markovian.
Electron-Phonon Coupling (The "Dancing Phonons"): Imagine the hallway floor isn't just bumpy; it's made of trampoline springs that vibrate and dance. When an electron hits a spring, the spring wiggles, absorbs some energy, and then wiggles back, pushing the electron again later.
- The Paper's Finding: This is Scenario B. Because the springs (phonons) take time to vibrate and settle, the electron feels the effect of the collision for a long time. It has a "long memory." The paper calls this Non-Markovian.

The Detective Tool: The "Spectral Function"

How do we know if an electron has memory if we can't see it? The author uses a mathematical tool called the Spectral Function.

Think of the Spectral Function as a sound wave recorder.

If the electron has no memory (Static Disorder), the sound wave dies out immediately. It's a sharp, short click.
If the electron has memory (Phonons), the sound wave rings out like a bell. It oscillates (wiggles back and forth) and fades away slowly.

The paper argues that by looking at this "ringing" pattern in the data, scientists can diagnose whether a system is behaving with memory or without it, even without watching the electron move in real-time.

The "Self-Consistent" Twist

The paper also compares two ways of doing the math:

The "First Guess" (Born Approximation): You calculate the effect of the collision once, assuming the electron is a simple, perfect particle.
The "Second Guess" (Self-Consistent Born): You realize the electron gets messy and slows down after the first collision, so you recalculate the effect taking that messiness into account.

The Discovery:

For the Static Noise, it doesn't matter which method you use. The electron still forgets instantly. The math stays simple.
For the Dancing Springs (Phonons), the "Second Guess" changes everything. When you account for the electron getting messy, the "memory" of the collision actually gets shorter and more localized. The electron starts to forget faster than you thought. This suggests that strong interactions can actually make a "memory-heavy" system start to look more like a "no-memory" system.

The Final Test: Two Different Hallways

To prove this isn't just a fluke of one specific material, the author tested two very different types of 2D grids:

The Hofstadter Model: A grid with a magnetic field that makes the electrons' paths twist and turn in complex patterns (like a maze).
The RKKY Model: A grid where atoms talk to each other over long distances (like a long-distance phone call).

The Result:
Even though these two grids are totally different, the rule held true:

Static bumps always led to "no memory" behavior.
Vibrating springs always led to "memory" behavior.

This proves that the type of memory depends on how the electron interacts (static vs. vibrating), not on the specific shape of the material it's moving through.

Summary of the Conclusion

The paper builds a unified bridge between three things:

Microscopic Physics: What happens when an electron hits a bump or a spring.
Mathematical Structure: How the equations (Green's functions) show time delays.
Observable Results: How the "memory" shows up in the transmission of electricity.

The Takeaway:
If you want to know if a tiny electronic system has "memory," don't just look at the electrons; look at the environment they are in. If the environment is static, the system forgets instantly. If the environment vibrates (like phonons), the system remembers, and this memory shows up as a specific "ringing" signature in the electrical current. The author provides a toolkit to spot these signatures in future experiments.

Technical Summary: Reinterpreting Memory Effects in Nonequilibrium Systems

Problem Statement
Quantum transport in low-dimensional and nanoscale systems is fundamentally governed by interactions, coherence, and coupling to external environments. While many transport approaches rely on the Markovian approximation—assuming the system instantaneously loses information about its past—this assumption often fails in interacting or energy-exchange-dominated regimes. The central problem addressed is the lack of a unified framework connecting microscopic scattering mechanisms (elastic vs. inelastic) to the emergence of Markovian versus non-Markovian dynamics, and how these distinct memory signatures manifest in observable steady-state transport properties. Specifically, the paper seeks to clarify how temporal nonlocality in self-energies, arising from different scattering sources, translates into measurable spectral and transmission signatures.

Methodology
The authors employ the Non-Equilibrium Green's Function (NEGF) formalism within the Schwinger-Keldysh non-equilibrium quantum field theory framework. The study proceeds through several theoretical and numerical stages:

Formal Foundation: Starting from a 2D tight-binding Hamiltonian, the Dyson equation is formulated on the Keldysh contour. The Kadanoff-Baym equations are derived to analyze the temporal convolution structure of the self-energy, which encodes memory effects.
Scattering Mechanisms: The paper contrasts two primary scattering mechanisms:
- Elastic Scattering: Modeled via static disorder (impurities), leading to time-local self-energies.
- Inelastic Scattering: Modeled via electron-phonon coupling, leading to temporally nonlocal self-energies.
Approximations: The study compares the second-order Born approximation (using bare propagators) with the Self-Consistent Born Approximation (SCBA, using dressed propagators) to evaluate how feedback loops influence memory kernels.
Renormalization Group (RG) Perspective: The authors utilize 1PI (one-particle irreducible) and 2PI (two-particle irreducible) effective actions to interpret memory effects through the lens of coarse-graining.
Model Systems: The theoretical framework is applied to two paradigmatic 2D models: the Hofstadter model (magnetic flux-induced complex hopping) and an RKKY-coupled system (long-range oscillatory interactions).
Steady-State Analysis: The study transitions from time-domain dynamics to the frequency domain to analyze steady-state transmission functions ( $T(E)$ ) and spectral functions ( $A(E)$ ) as diagnostic probes.

Key Contributions and Results

Distinct Memory Signatures in Spectral Functions: The paper demonstrates that elastic scattering (static disorder) produces a rapidly decaying memory kernel, characteristic of Markovian dynamics. In contrast, electron-phonon coupling generates temporally nonlocal self-energies with oscillatory, long-lived decay, signifying genuine non-Markovian behavior. These differences are directly reflected in the spectral function, proposed here as a diagnostic probe.
Role of Self-Consistency (Born vs. SCBA): A comparison between the Born and SCBA approximations reveals that while elastic scattering remains strictly local in time regardless of self-consistency, inelastic scattering exhibits significant renormalization. In the SCBA, self-consistent quasiparticle damping progressively localizes the memory kernel near $\tau=0$ , suppressing long-time temporal correlations and driving the system toward effectively Markovian dynamics at stronger coupling strengths.
RG Interpretation of Memory: Through the analysis of effective actions, the authors show that stronger coarse-graining (associated with 1PI effective actions) naturally suppresses memory, favoring Markovian dynamics. Conversely, weaker coarse-graining (associated with 2PI effective actions) retains temporal nonlocality and memory kernels, preserving non-Markovian features.
Steady-State Signatures in Transmission: The study establishes that non-Markovian effects in the time domain manifest in the steady-state energy domain as frequency-dependent self-energies. This results in nontrivial renormalization profiles in the transmission function $T(E)$ , such as peak splitting and frequency-dependent broadening, distinct from the simple monotonic suppression seen in elastic (Markovian) cases.
Universality: These qualitative distinctions between elastic and inelastic scattering hold true across different lattice complexities, persisting in both the topologically complex Hofstadter model and the long-range RKKY system. This suggests that the observed signatures originate from the structure of the scattering kernel rather than specific lattice details.

Significance and Claims
The paper claims to provide a unified connection between scattering mechanisms, memory effects, and quantum transport in low-dimensional systems. By linking microscopic theory, diagrammatic structure, and renormalization-group interpretations, the work clarifies the physical origin of memory effects.

The authors assert that the spectral function and transmission function serve as experimentally accessible probes for identifying non-Markovian behavior without relying on explicit time-resolved dynamics. The study emphasizes that while existing literature has addressed driven transport or open quantum systems separately, this work systematically explores the role of scattering mechanisms in controlling memory effects and their manifestation in steady-state transport quantities.

The paper concludes modestly, noting that while it establishes the theoretical framework and steady-state signatures, the specific role of external driving in controlling these memory effects and their detailed manifestation in driven steady-state transport remains a subject for future investigation.